statistical approaches for identification and

Environment Protection Engineering

Vol. 45 2019 No. 4

DOI: 10.37190/epe190402

LINHUA SUN1, 2

STATISTICAL APPROACHES FOR IDENTIFICATION

AND QUANTIFICATION OF SOIL TRACE ELEMENTAL

POLLUTION NEAR THE TRAFFIC WAY

The concentrations of eight trace elements (As, Co, Cr, Cu, Fe, Mn, Pb and Zn) in the soil near

a traffic way of Suzhou, Anhui province, China have been determined for the pollution assessment and

source identification (along with quantification). The results indicate that Fe is the most abundant ele-

ment followed by Mn, Zn, Cr, Pb, Cu, Co and As. These elements have low-medium coefficients of

variation (0.059–0.293), indicating that some of them might have multisources. The combination of

single pollution, geoaccumulation and the Nemerow composite indices suggest that the soils in this

study are slightly polluted. Based on multivariate statistical analyses (including correlation, cluster and

factor analyses), three sources responsible for the trace elemental concentrations in the soils have been

identified, including geogenic, agricultural and traffic-related sources. Their mean contributions calcu-

lated based on the Unmix model are 35.9, 31.8 and 32.4%, respectively.

1. INTRODUCTION

With the vigorous development of China’s national economy, the pace of road con-

struction is growing. According to the Statistical Bulletin on Transportation Industry

Development in 2016, China, China’s total traffic way mileage has reached 4.70 million

kilometres at the end of 2016. Along with this situation, the environmental problems

(e.g., noise, vibration, emission of gases, water, residue and waste) related to the traffic

have attracted a lot of attention. And therefore, a large number of studies have been

carried out for the solution of these environmental problems and most of them focused

on the noise and air pollution [1–4].

_________________________

1School of Resources and Civil Engineering, Suzhou University, Anhui 234000, China, e-mail ad-

dress: [email protected] 2Key Laboratory of Mine Water Resource Utilization of Anhui Higher Education Institute, Suzhou

University, Anhui 234000, China.

22 L. SUN

Because of their harmful effects for human beings, some of the trace elements, es-

pecially the toxic ones (e.g., Hg, Cd, Pb, Zn), have long been concerned by scientists,

and a large number of studies related to their toxicological characteristics [5, 6], con-

centrations, distribution and forms in the environments [7, 8], sources [9], migration,

enrichment and transformation [10] have been carried out. With the development of the

traffic way of China, the vehicle exhaust emissions have become an important factor for

the increasing concentration of trace elements such as Pb, Cu, Zn and Cd in the atmos-

phere [11], and which can increase the trace elemental concentrations in soil on both

sides of traffic way (mostly, Pb, Cu, Zn, Cd and Mn) [12–16].

In this study, soil samples have been collected from the side of a traffic way with

high density of traffic in the Suzhou city, Anhui province, China, and the concentrations

of As, Co, Cr, Cu, Fe, Mn, Pb and Zn have been measured for the pollution assessment

and source approximation (especially the identification and quantification of the traffic-

related pollution of trace elements). The study can provide information for the environ-

mental management of the study area.

2. MATERIALS AND METHODS

2.1. THE STUDY AREA

Suzhou is the north gate of the Anhui province, China. It is located at the south of the

Huang-Huai plain, adjacent to Xuzhou of Jiangsu and Heze of Shandong in the north, Yong-

cheng of Henan in the west, Suqian of Jiangsu in the east. There are many rivers in the area,

including the Kui, Sui, Tuo and Hui Rivers, all of them flow from northwest to southeast,

and end in the Huai River or the Hongze lake. The annual precipitation is 857 mm, with an

average temperature of 14.4 degrees (centigrade). Agriculture and coal related industries are

the most important industries in the area. The main crops in the area include wheat, corn,

soybean, cotton, potato, rapeseed, peanuts and fruit etc. And therefore, the quality of soil is

important for the development of the area.

During the past ten years, the economy of Suzhou has undergone rapid develop-

ment, especially the development of transportation, including the railways and traffic

ways. And now, people in Suzhou can arrive in Beijing and Shanghai in three hours,

and Xi’an and Wuhan in four hours. In relation to this high speed of development, many

kinds of environmental problems such as air, water and soil pollution occurred and a se-

ries of studies have been carried out [17–19].

2.2. SAMPLING AND ANALYSIS

The Suzhou Avenue is a urban express connecting the urban area and the high rail-

way east station of Suzhou, and it is a traffic way with high density of traffic (Fig. 1). It

Statistical approaches for identification of soil trace elemental pollution 23

is an eight-lane road with two non-motorized lanes, the width is 40 m, with an average

traffic volume of near 10 000 vehicles per day and a peak period of nearly 1500 vehicles

per hour. A total of 30 surface soil samples (less than 10 cm deep) near the traffic way

have been collected in March 2019. The detailed sampling distributions are shown in

Fig. 1. All of the samples were collected at the southern edge of the traffic way (within

1 m from the roadside). The distance between each sampling point was near 150 m.

Fig. 1. Location of the study area and sample distributions

After collection, all samples were air-dried in natural conditions, and the debris of ani-

mals and plants were removed by hands. Then the samples were powdered to 200 meshes

(<0.075 mm) after parching for 24 h at 80 °C in a dryer. Samples were tableted using

a 30 t condenser, and then analyzed by XRF (Innov-X Explorer 9000 SDD, USA) for

measuring the concentrations of trace elements such as As, Co, Cr, Cu, Fe, Mn, Pb and

Zn in the Key Laboratory of Mine Water Resource Utilization of Anhui Higher Educa-

tion Institute, Suzhou University, Anhui Province, China. National standard sediment

sample of China (GSS-16) was analyzed simultaneously for calibration (once per 10

samples). And the final concentrations of the trace elements were calculated from:

t sm

m

C SC

S= (1)

where Cm is the concentration of a sample, Ct is its concentration recorded with the

instrument, Ss and Sm are the standard and measured concentrations of the standard sam-

ples (GSS-16), respectively.

2.3. DATA TREATMENT PROCEDURES

All of the data were firstly processed for statistical analysis by the Mystat 12 soft-

ware, and the minimum, maximum, mean, and standard deviation, the coefficient of

24 L. SUN

variation and the p-value of the normal distribution test have been obtained. And then,

a series of parameters, including the single pollution index Pi [20], the geoaccumulation

index Igeo [21] and the Nemerow composite index Ps [22] have been determined for the

pollution assessment of the soil samples. Finally, the statistical analyses (including the

correlation, cluster and factor analyses) [23] have been performed for information on

the sources of the pollution, and then the Unmix model provided by the US Environ-

mental Protection Agency (EPA) [24] has been used to obtain the quantitative infor-

mation about the source of them.

3. RESULTS AND DISCUSSION

3.1. TRACE ELEMENTAL CONCENTRATIONS

The concentrations of the trace elements in 30 soil sampled are given in Table 1. As

can be seen from the table, iron is the element with the highest mean concentration

(22,987–30,889 mg/kg, mean 26,689 mg/kg), and then followed by the Mn, Zn, Cr, Pb, Cu,

Co and As, their mean concentrations are 389, 119, 49.6, 35.9, 24.2, 12.1 and 9.96 mg/kg,

respectively.

T a b l e 1

Descriptive statistics of trace elemental concentrations in the soil samples

Species As Co Cr Cu Fe Mn Pb Zn

Minimum, mg/kg 7.82 9.00 31.3 15.3 22,987 328 27.9 66.3

Maximum, mg/kg 14.0 12.9 68.7 43.4 30,889 497 50.3 191

Mean, mg/kg 9.96 12.1 49.6 24.2 26,689 389 35.9 119

Standard deviation, mg/kg 1.82 0.721 7.17 6.13 2,028 33.5 6.01 34.9

Coefficient of variation 0.183 0.059 0.145 0.254 0.076 0.086 0.167 0.293

p-value <0.01 <0.01 >0.15 0.144 >0.15 0.113 0.034 >0.15

Mean Pi 0.889 0.957 0.813 1.07 0.908 0.667 1.38 1.61

Highest Pi 1.25 1.01 1.13 1.92 1.05 0.853 1.93 2.58

Igeo 0.162 0.150 0.091 0.118 0.035 0.058 0.108 0.080

Highest Igeo 0.170 0.160 0.097 0.127 0.035 0.059 0.112 0.086

Background 11.2 12.7 61.0 22.6 29 400 583 26.0 74.2

The background data are from the CEPA [26].

The coefficient of variation (CV – standard deviation/mean) is an index showing the

degree of variability with respect to the mean of the population. A high CV (>0.90)

means a high degree of spatial variation and a high degree of anthropogenic contribution

(point pollution), whereas a low CV (<0.10) means low degree of spatial variation and

low degree of anthropogenic contribution or surface pollution [25]. In this study, Co, Fe

and Mn have low CVs, which means that the concentrations of these elements vary


slightly from sample to sample, indicating that they have undergone low degree or no anthropogenic influence. As to other elements, they have a medium CVs values between 0.145 and 0.293, which indicates a moderate spatial inhomogeneity with moderate an-thropogenic contribution. Such results can also be achieved from the p-values of the normal distribution test. As can be seen from the table, most of the elements except for As, Co and Pb have p-values higher than 0.05, implying that they can pass the normal distribution test, which may also suggest that the elements except for As, Co and Pb might have single source.

3.2. POLLUTION ASSESSMENT

Previous studies revealed that the single pollution index Pi is a good indicator for monitoring the degree of pollution

mi

s

CPC

(2)

Cm and Cs are the concentrations of the sample and background, respectively. Pi < 1 means light pollution, Pi between 1 and 3 means moderate pollution, and Pi > 3 means considerable pollution [20].

The soil environmental background values of China [26] were chosen to be the Cs values, and the calculated mean Pi values are listed in Table 1. The results indicate that the soils in this study are moderately polluted with Cu, Pb and Zn, (Pi equal to 1.07, 1.38 and 1.61, respectively), whereas for other elements Pi < 1. Despite the fact that most of the average concentrations of the trace elements with light pollution (As, Co, Cr and Fe) in this study are lower than the soil environmental background values, there are differences between the samples from different locations: it can be seen from the table that the maximum concentrations of As, Co, Cr and Fe (14.0, 12.9, 68.7 and 30 889 mg/kg, respectively) are higher than those of the background (11.2, 12.7, 61.0 and 29 400 mg/kg, respectively), and the highest Pi values are 1.25, 1.01, 1.13 and 1.05, respectively, im-plying that the distribution of trace elements in the study area is heterogeneous.

The geoaccumulation index Igeo

geo 2log1.5

m

s

CIC

(3)

enables the assessment of contamination degrees by comparing the current and prein-dustrial concentrations [21]. There are no local preindustrial concentrations of trace el-ements, and therefore, the soil environmental background values [26] were chosen as the Cs values. The values of Igeo < 0 correspond to unpolluted samples, those between

26 L. SUN

0 and 1 light pollution, 1–3 moderate pollution, 3–5 heavy pollution, and > 5 serious

pollution [21]. The calculated Igeo values listed in Table 1 point to light pollution with

all the elements (0 < Igeo < 1).

The Nemerow composite index Ps method takes into account all the individual eval-

uation factors:

( ) ( )

22 2

1/

2

i i

s

Pm PxP

+=

(4)

where Pim is the average of single pollution index of all elements, and Pix is the maxi-

mum value of the single pollution index of all elements. Ps < 0.7 means safety domain,

0.7 ≤ Ps < 1.0 precaution domain, 1.0 ≤ Ps < 2.0 slightly polluted domain, 2.0 ≤ Ps < 3.0

moderately polluted domain, and Ps > 3.0 seriously polluted domain [22]. In this study,

the calculated Ps values for all of the samples range from 0.98 to 2.01 (mean 1.40), only

one sample has 0.7 ≤ Ps < 1.0 and another Ps > 2.0, which means that most of the soil

samples in this study can be classified to be slightly polluted.

3.3. STATISTICAL ANALYSES FOR SOURCE IDENTIFICATION AND QUANTIFICATION

Correlation analysis. The correlation analysis between species is useful as it can

point out associations between variables that can show the overall coherence of the da-

taset, and thus indicate the participation of the individual chemical parameters in several

influence factors [27]. The results of the correlation analysis are shown in Table 2. As

can be seen from the table, close positive relationships have been identified between the

following elements: Co–Mn, Cr–Cu–Fe–Mn, Cu–Fe, Fe–Mn and Pb–Zn. Such results

suggest that these elemental pairs might have similar sources or, have been affected by

similar factors. However, Pb and Zn show negative correlations with almost all of the

other elements except for Co, implying that they have different sources.

T a b l e 2

Results of the correlation analysis

As Co Cr Cu Fe Mn Pb

Co –0.214

Cr 0.260 0.291

Cu 0.242 0.213 0.476a

Fe 0.176 0.277 0.537a 0.423a

Mn 0.005 0.421a 0.529a 0.068 0.521a

Pb 0.080 0.214 –0.319 –0.178 –0.643a –0.194

Zn –0.186 0.295 –0.330 –0.121 –0.598a –0.052 0.791

aMeans significant at a = 0.05, ra = 0.361.


Cluster analysis. Cluster analysis comprises of a series of multivariate methods

which are used to find true groups of data. In clustering, the objects are grouped such

that similar objects fall into the same class. The method has long been used for environ-

mental studies [13]. In this study, the hierarchical R-mode cluster analysis has been ap-

plied to the data, and the “Ward” linkage and the “Pearson” distance have been chosen

for calculation, and the results are shown in Fig. 2 as a dendrogram. As can be seen from

the figure, two main groups can be identified: Pb–Zn (group 1) and As–Co–Cr–Cu–Fe–Mn

(group 2), which indicate that the elements in the similar group might have similar

sources.

Fig. 2. Dendrogram of R-mode cluster analysis

Factor analysis. Factor analysis is a commonly used statistical method for classifi-

cation, simplification of the data and finding the most important variables in the dataset.

In geochemical studies, factor analysis has long been used for tracing elemental sources

[28]. In this study, based on the criterion of initial eigenvalue higher than one, three

factors have been obtained based on the factor analysis (Mystat version 12.0) (Table 3).

According to previous studies [29], factor loadings can be classified as strong, me-

dium and weak with the values of >0.75, 0.75–0.50 and 0.50–0.30, respectively. As can

be seen from the table, the first factor, which accounts for 38.7% of the total variance

explanation, has moderate negative loadings of Fe and strong positive loadings of Pb

and Zn. As to factor 2, it accounts for 22.6% of the total variance explanation and has

strong positive loadings of Co and Mn and moderate positive loadings of Cr and Fe.

The third factor has 15.5% of the total variance explanation and has strong positive

loadings of As, and moderate positive loadings of Cr and Cu. In combination with pre-

vious studies, the first factor with high Pb and Zn loadings can be explained to be the

Pb–Zn related sources (e.g., the traffic-related, because Pb and Zn are the most abundant

elements released by vehicle exhaust emissions and wearing of tyres) [25]. The second

factor can be explained to be the geogenic source because the elements with high load-

ings in this factor are mostly geogenic (e.g., Co, Cr, Fe and Mn). The third factor with

28 L. SUN

high loadings of As and Cu, to a lesser extent, Cr, might be an indication of agricultural

activities, especially the pesticides and fertilizers (because As and Cu are the most pop-

ular elements applied for pesticides production).

T a b l e 3

Results of factor analysis

Element Factor 1 Factor 2 Factor 3

As –0.01 –0.23 0.85

Co 0.30 0.84 –0.04

Cr –0.33 0.58 0.53

Cu –0.11 0.30 0.68

Fe –0.69 0.53 0.29

Mn –0.20 0.78 0.03

Pb 0.95 –0.07 0.04

Zn 0.92 0.12 –0.16

Eigen value 3.10 1.81 1.24

Variance

explained, % 38.7 22.6 15.5

Unmix model analysis. The Unmix model is a mathematical receptor model used

for quantifying the sources of contaminants contributing to sediment, water and air sam-

ples [24]. It is based on the reducing of a large number of variables in complex analytical

data sets to combinations of species called source types and source contributions. The

source types are identified by comparing them to measured profiles, whereas the source

contributions are used to determine how much each source contributed to a sample [30].

Based on the calculation, three sources have been identified and the results are listed in

Table 4 and shown in Fig. 3.

T a b l e 4

Source profiles [mg/kg] and proportions [%]

Element Source 1 Source 2 Source 3 Proportion 1 Proportion 2 Proportion 3

As 2.29 1.99 5.72 22.9 19.9 57.2

Co 2.68 3.40 6.05 22.1 28.0 49.9

Cr 14.3 9.33 26.0 28.8 18.8 52.4

Cu 11.8 5.07 7.29 48.8 21.0 30.2

Fe 6810 5570 14200 25.6 21.0 53.4

Mn 83.4 95.7 209 21.5 24.7 53.9

Pb 4.77 14.6 16.5 13.3 40.7 46.0

Zn 12.7 67.5 39.1 10.6 56.6 32.8

These three sources have Min Rsq = 0.98 and Min Sig/Noise = 2.17, higher than the

minimum requirement of the model (Min Rsq > 0.8 and Min Sig/Noise > 2), suggesting


that the modeling is efficient [24]. As can be seen from the table, the first source is

characterized by highest Cu content (48.8%) relative to other sources, and then followed

by Cr (28.8%), Fe (25.6%), As (22.9%), Co (22.1%), Mn (21.5%), Pb (13.3%) and Zn

(10.6%), whereas the second source is characterized by the highest contents of Zn

(56.6%) relative to other sources, and then followed by the Pb (40.7%), Co (28.0%),

Mn (24.7%), Cu and Fe (21.0%), As (19.9%) and Cr (18.8%). The third source is char-

acterized by the highest proportions of most of the elements (As (57.2%), Mn (53.9%),

Fe (53.4%), Cr (52.4%), Co (49.9%) and Pb (46.0%)), and then followed by the Zn

(32.8%) and Cu (30.2%).

Fig. 3. Variations of source contributions

In consideration with the natural condition of the study area, there are three sources

responsible for the soil trace elements in this study: the geogenic source related to the

formation of the soil, the agricultural activities (application of fertilizers and pesticides)

and traffic-related activities (vehicle exhaust emissions and wearing of tyres). In com-

bination with the above analyses, the Source 3 can be explained to be the geogenic

source, because most of the elements including Fe and Mn (mostly, geogenic) are con-

tributed by this source. The contribution of this source for the elemental concentrations

of all of the samples range from 0 to 68.0% (mean 35.9%). Comparatively, the source 1 can

be explained to be the source related to the agricultural activities because this source is

30 L. SUN

characterized by highest content of Cu, and then followed by As and others, which are

always related to the productions of pesticides. The contribution of this source for the

elemental concentrations of all of the samples range from 2.40 to 81.2% (mean 31.8%).

The source 2 should be explained to be the traffic related source as it is characterized

by high contents of Zn and Pb, and its contributions are 0 to 73.4% (mean 32.4%).

4. CONCLUSIONS

• The elemental concentrations are Fe > Mn > Zn > Cr > Pb > Cu > Co > As. They

have low-medium coefficients of variation (0.059–0.293).

• The single pollution, geoaccumulation and the Nemerow composite indices sug-

gest that the soils in this study are slightly polluted.

• Statistical analyses indicate that three sources are responsible for the soil ele-

mental concentrations, the geogenic, agricultural and traffic-related sources, and their

mean contributions calculated by the Unmix model are 35.9%, 31.8% and 32.4%, re-

spectively.

ACKNOWLEDGEMENTS

The authors are very grateful to the students Han Sun, Zixiang Zhao, Yu Li and Xiaojun Gao for their

help in sampling and analyses. Thanks to the anonymous reviewer for his/her valuable comments. This

work was financially supported by the Academic Funding for Top-talents in Disciplines of Universities in

Anhui Province (gxbjZD48).

REFERENCES

[1] CARSLAW D.C., Evidence of an increasing NO2/NOx emissions ratio from road traffic emissions, Atmos.

Environ., 2005, 39 (26), 4793.

[2] ZECHMEISTER H.G., HOHENWALLNER D., RISS A., HANUS-ILLNAR A., Estimation of element deposition

derived from road traffic sources by using mosses, Environ. Pollut., 2005, 138 (2), 238.

[3] MEHDI M.R., KIM M., SEONG J.C., ARSALAN M.H., Spatio-temporal patterns of road traffic noise pollu-

tion in Karachi, Pakistan, Environ. Int., 2011, 37 (1), 97.

[4] DE SILVA S., BALL A.S., HUYNH T., REICHMAN S.M., Metal accumulation in roadside soil in Melbourne,

Australia. Effect of road age, traffic density and vehicular speed, Environ. Pollut., 2016, 208, 102.

[5] JI C., CAO L., LI F., Toxicological evaluation of two pedigrees of clam Ruditapes philippinarum as bio-

indicators of heavy metal contaminants using metabolomics, Environ. Toxicol. Pharmacol., 2015, 39

(2), 545.

[6] AROJOJOYE O.A., OYAGBEMI A.A., AFOLABI J.M., Toxicological assessment of heavy metal bioaccumu-

lation and oxidative stress biomarkers in Clarias gariepinus from Igbokoda River of South Western

Nigeria, Bull. Environ. Contam. Toxicol., 2018, 100, 765.

[7] ISLAM M.S., AHMED M.K., RAKNUZZAMAN M., HABIBULLAH-AL-MAMUN M., ISLAM M.K., Heavy metal

pollution in surface water and sediment: a preliminary assessment of an urban river in a developing

country, Ecol. Indic., 2015, 48, 282.


[8] TÓTH G., HERMANN T., DA SILVA M.R., MONTANARELLA L., Heavy metals in agricultural soils of the European Union with implications for food safety, Environ. Int., 2016, 88, 299.

[9] HUANG Y., LI T., WU C., HE Z., JAPENGA J., DENG M., YANG X., An integrated approach to assess heavy metal source apportionment in peri-urban agricultural soils, J. Hazard. Mater., 2015, 299, 540.

[10] LI R., ZHAO W., LI Y., WANG W., ZHU X., Heavy metal removal and speciation transformation through the calcination treatment of phosphorus-enriched sewage sludge ash, J. Hazard. Mater., 2015, 283, 423.

[11] WALRAVEN N., VAN OS B.J.H., KLAVER G.T., MIDDELBURG J.J., DAVIES G.R., The lead (Pb) isotope signature, behaviour and fate of traffic-related lead pollution in roadside soils in The Netherlands, Sci. Total Environ., 2014, 472, 888.

[12] CHEN X., XIA X., ZHAO Y., ZHANG P., Heavy metal concentrations in roadside soils and correlation with urban traffic in Beijing, China, J. Hazard. Mater., 2010, 181 (1–3), 640.

[13] ZHANG H., WANG Z., ZHANG Y., DING M., LI L., Identification of traffic-related metals and the effects of different environments on their enrichment in roadside soils along the Qinghai–Tibet highway, Sci. Total Environ., 2015, 521, 160.

[14] ZUPANČIČ N., Lead contamination in the roadside soils of Slovenia, Environ. Geochem. Health., 1999, 21 (1), 37.

[15] LINDE M., BENGTSSON H., ÖBORN I., Concentrations and pools of heavy metals in urban soils in Stock-holm, Sweden, Water, Air, Soil Pollut. (Focus), 2001, 1 (3–4), 83.

[16] GUO G.H., CHEN T.B., SONG B., YANG J., HUANG Z.C., LEI M., CHEN Y.C., Emissions of heavy metals from road traffic and effect of emitted lead on land contamination in China. A primary study, Geogr. Res., 2007, 26 (5), 922.

[17] MEI J., LI Z., SUN L., GUI H., WANG X., Assessment of heavy metals in the urban river sediments in Suzhou City, Northern Anhui Province, China, Proc. Environ. Sci., 2011, 10, 2547.

[18] SUN L., Lead pollution in response to transportation: a case study in the rural-urban fringe zone of Suzhou, northern Anhui Province, China, J. Chem. Pharm. Res., 2014, 6 (7), 2370.

[19] SUN L., LIU X., CHENG C., Quality evaluation of water from subsidence area and controlling factor analysis. Zhuxianzhuang case study, Nat. Environ. Pollut. Technol., 2016, 15 (3), 1035.

[20] LIANG J., CHEN C., SONG X., HAN Y., LIANG Z., Assessment of heavy metal pollution in soil and plants from Dunhua sewage irrigation area, Int. J. Electrochem. Sci., 2011, 6 (11), 5314.

[21] PRAVEENA S.M., AHMED A., RADOJEVIC M., ABDULLAH M.H., ARIS A.Z., Heavy metals in mangrove sur-face sediment of Mengkabong Lagoon, Sabah. multivariate and geoaccumulation index approaches, Int. J. Environ. Res., 2008, 2 (2), 139.

[22] DAI J., LI S., ZHANG Y., WANG R., YU Y., Distributions, sources and risk assessment of polycyclic aromatic hydrocarbons (PAHs) in topsoil at Jinan city, China, Environ. Monit. Assess., 2008, 147 (1–3), 317.

[23] HAN Y., CAO J., POSMENTIER E.S., Multivariate analysis of heavy metal contamination in urban dusts of Xian, Central China, Sci. Total Environ., 2006, 355 (1–3), 176.

[24] NORRIS G., VEDANTHAM R., DUVALL R., HENRY R.C., EPA Unmix 6.0 fundamentals & user guide, US Environmental Protection Agency, Office of Research and Development, Washington, DC, 2007.

[25] SARKAR D., DATTA R., HANNIGAN R., Concepts and Applications in Environmental Geochemistry, Vol. 5, Elsevier, 2011.

[26] CEPA, Elemental background values of soils in China, Chinese Environmental Protection Admin-istration, Environmental Science Press of China, Beijing 1990.

[27] COBELO-GARCÍA A., PREGO R., Influence of point sources on trace metal contamination and distribu-tion in a semi-enclosed industrial embayment: the Ferrol Ria (NW Spain), Estuar. Coast. Shelf. S., 2004, 60 (4), 695.

[28] MAIZ I., ARAMBARRI I., GARCIA R., MILLAN E., Evaluation of heavy metal availability in polluted soils by two sequential extraction procedures using factor analysis, Environ. Pollut., 2000, 110 (1), 3.

32 L. SUN

[29] SUN L.H., GUI H.R., Establishment of water source discrimination model in coal mine by using hydro-

geochemistry and statistical analysis. A case study from Renlou Coal Mine in northern Anhui Prov-

ince, China, J. Coal. Sci. Eng. (China), 2012, 18 (4), 385.

[30] SUN L., PENG W., CHENG C., Source estimating of heavy metals in shallow groundwater based on

UNMIX Model. A case study, Indian. J. Geo-Mar. Sci., 2016, 45 (6), 756.

statistical approaches for identification and

Documents